JMIR Publications

ABSTRACT

Background:

Rapid technological advance has given rise to the emergence of generative artificial intelligence (GenAI). GenAI tools such as ChatGPT and DALL-E can generate text, images, or other media in response to prompts and have permeated various facets of daily life. The current literature reports mixed relationships between GenAI use and life satisfaction.

Objective:

By synthesizing simplified Technology Acceptance Model and self-determination theory, this study aimed to examine the associations between GenAI use and life satisfaction, the role of perceived AI usefulness, behavioral intention to use AI, and AI literacy, and explore potential heterogeneity across gender and age subgroups.

Methods:

A representative sample of 1,800 community-dwelling adults (mean age = 49.3 years, SD = 14.9, 55.3% female) were recruited in a population-based survey via a two-stage random sampling design in Hong Kong in spring 2024. Participants completed self-report measures assessing perceived AI usefulness, behavioral intention to use AI, GenAI use, AI literacy, and life satisfaction. Multivariate analysis of covariance was conducted to examine potential gender and age differences in the study variables. Structural equation modelling (SEM) was used to examined the associations between AI-related constructs, GenAI use, and life satisfaction in the whole sample and across socio-demographic subgroups.

Results:

Two-fifths (38.5%) of the sample reported GenAI use in the past year. There were significant age differences with declining trends in perceived AI usefulness, behavioral intention to use AI, GenAI use, and AI literacy from young adults to older adults. Respondents with more than 2 hours of GenAI use daily reported significantly higher perceived AI usefulness, behavioral intention to use AI, AI literacy, and life satisfaction than non-users. In the SEM, the direct effect of GenAI use on life satisfaction was not significant. There were significant and positive indirect effects (αβ = 0.186, 95% CI = 0.134 – 0.242) from perceived AI usefulness to life satisfaction via behavioral intention to use AI, GenAI use, and AI literacy. In subgroup analyses, the total indirect effects were stronger in males (αβ = 0.260, 95% CI = 0.177 – 0.348) and older adults (αβ = 0.227, 95% CI = 0.113 – 0.367) than females (αβ = 0.112, 95% CI = 0.055 – 0.180) and young adults (αβ = 0.124, 95% CI = 0.038 – 0.214), respectively.

Conclusions:

The present findings synthesize TAM and self-determination theory by demonstrating a mediating role of AI literacy in the relationship between behavioral intention to use AI, GenAI use and life satisfaction in Hong Kong adults. The sample heterogeneity in the indirect effects across gender and age groups have practical implications on the potential use of GenAI to improve population health among males and older adults.